Subjects

Volunteers were selected among the farmers belonging to rural areas of Tehsil Burewala (Vehari district) and Jahanian (Khanewal district) in Punjab (Pakistan). In total, 59 male exclusive smokers, with age ranging 20–80 years (mean = 58.8 ± 14.7) and 8–65 years of smoking (mean = 37.7 ± 16.8), participated in this study. The smokers used the hookah in 1–8 daily sessions in the form of 1–4 daily preparations and occasional smokers were excluded. About 1–2 "chattaks" (60 to 120 g; 1 chattak = 58.125 g) of a Desi Punjab tobacco-molasses mixture were used in each preparation. The approximate quantity of tobacco packed inside the chillum (bowl) was almost same in all the subjects. Subjects were divided into 3 groups according to their daily number of preparations, sessions and total smoking time: Light (1; 1; ≤ 20 minutes); Medium (1–3; 1–3; >20 min to ≤ 2 hrs) and Heavy smokers (2–4; 3–8; >2 hrs to ≤ 6 hrs). Before taking blood samples, the subjects were interviewed to make sure that they did not use any other tobacco or smoking product (cigarette, bidis, chewed tobacco). They were clinically examined to exclude any disease (which might influence CEA levels) or narcotic habit and questioned about their smoking career (Figures 2 and 3).

Materials

Desi Punjab air-dried and sun-cured tobacco leaves were used to prepare a tobacco-molasses mixture. The molasses element is a dense brownish sap formed during purification of cane sugar (extracted by boiling). The final mixture contains molasses and tobacco in 1:1 ratio by weight [19]. It is similar, to a certain extent, to Arabian jurâk, which contains pulpy fruit apart from tobacco and molasses. Such a mixture is cooked and 15 minutes of its smoking "would provide approximately the same amount of tobacco smoke as one cigarette" [20]. Mixing tobacco with molasses is a very ancient habit. A WHO report dates back "the addition of molasses to burley tobacco in the nineteenth century to create "American" blended tobacco" [21]. However, early health-oriented anthropological research on hookah smoking showed that it is much older and can be traced back to the 17^th century thanks to the early relation by an Arab traveller in India [22]. The smoking mixture is deposited above a small stone at the base of the Chilam (also spelled "chillum"), a funnel shaped 500–700 ml capacity container at the top of the hookah (Figure 4). Then, several pieces of glowing charcoal are put above. Notably, the tobacco-molasses mixture is not separated from the charcoal by an aluminium foil as is the case with the moassel (tobamel), the contemporaneous worldwide fashionable aromatic tobacco-molasses product, which also contains glycerol and flavouring essences [3]. Here, the charcoal is directly in contact with the tobacco-molasses mixture. The chilam is connected to a wooden pipe dipped in water in an airtight earthen base containing about 1 litre of water (Figure 5). A curved wooden pipe (Nari) of about 1 to 1.5 metre in length allows the user to inhale the smoke, which has bubbled through the liquid.

Methods

Carcinoembryonic antigen (CEA; a family of glycoproteins, MW~175000-2000 Daltons) was first identified in 1965 by Gold and Freedman in human colon cancer tissue extracts [23]. It is an important tumour marker for the diagnosis and monitoring of the therapy of diseases like colorectal cancer, lung cancer and ovarian cancer where elevated levels are observed. Levels are also raised in smokers but very few studies correlate CEA levels with smoking type and rate. The enhanced chemiluminescence immunometric technique was applied by using the Immulite-2000 system. This technique was introduced by Babson in 1991 [24]. It has gained worldwide acceptance because of its enhanced sensitivity, specificity and complete automation. The procedure principle involves a solid phase, two-site sequential chemiluminescent immunometric assay with two incubations each of 30 minutes. First incubation involves reaction of CEA with a polyclonal or monoclonal antibody coated on a plastic bead as a solid phase. Alkaline phosphatase-labelled antibody is then added to make an antigen antibody complex. Finally chemiluminescent substrate is added to estimate the reaction product by a luminometer. CEA immulite kits manufactured by DPC, USA were used for this purpose. The bead is housed in a proprietary test unit. This test unit serves as the reaction vessel for the immune reaction, the incubation and washing processes, and the signal development. The equipment used for automatic assay performance and measurement is automated immunoassay analyzer (immulite-2000) a continuous, random access instrument. A bar-code reader identifies tubes and reagent containers. After incubating the sample with the alkaline phosphatase reagent, the liquid reaction mixture in the test unit is rapidly separated from the bead when the bead is washed and the test unit is spun at a high speed on its vertical axis. The entire fluid contents (the sample, excess reagent, and wash solution) are transferred to a coaxial waste chamber, which is integral with the test unit. The bead is left with no residual, substrate. Following washing of the bead, addition of the substrate, a phosphate ester of adamantyl dioxetane, triggers the enzymatic reaction with bound alkaline phosphatase and creates the unstable adamantyl dioxetane anion. Breakdown of this unstable anion creates a prolonged 'glow' of light rather than a 'flash' of light and allows a longer window for numerous readings. The test units now enter a section of the instrument that excludes ambient light and allows for a 10-minute incubation for development of the luminescent signal. Light emission is detected by a photomultiplier tube (PMT). When each tube reaches the read station in front of the photomultiplier tube, photon counts are measured for 1 second through a 2°A neutral density filter, which attenuates the signal by a factor of 100. If counts are below a certain level, the attenuator is automatically removed. This automatic attenuation increases 100-fold the dynamic range of the photon multiplier tube. Counts are measured for 12 consecutive 1-second intervals. After discarding highest and lowest counts for each tube, the average counts per second are converted to analyte concentration by means of the stored calibration curves. The IMMULITE instrument has sufficient storage capacity to save data should any interruption of the communication with the external computer occurs and printed reports are generated for each sample by the system's computer.

5 cc of blood was taken from each smoker. Serum, obtained after clotting and centrifugation, was used in immunoassays for the determination of CEA levels. Samples were stored at 4°C in a fridge for future use.

Details about water changing, hookah size and actually used tobacco amount were precisely recorded during careful interviews with smokers. The protocol of blood sampling, blood handling and assay performance was reviewed by the medical board of Multan Institute of Nuclear Medicine and Radiotherapy (MINAR). It was considered exempt from Pakistani human subjects regulations because the analysis did not expose the volunteers to risk. There were no special ethical considerations in this study because no surgical and other interventions were involved. The safety of subjects was secured. All the subjects voluntarily donated their blood samples. The smokers were exhaustively and previously informed about the significance of the test and were assured of privacy regarding test values. Upon arrival at the local dispensary, participants reviewed study procedures with the medical staff and then completed blood sampling. All data were collected between August and December 2007.

Distribution curves were constructed from the data obtained on non-smokers and smokers (see Results section). Because of the nature of distribution of CEA levels among our individuals (it was not normal), Wilcoxon's rank sum two-sample test was applied to compare the variables using a program provided by the Institute of Phonetic Sciences of the University of Amsterdam (The Netherlands)[26].

General health aspects of hookah smoking

Health effects of hookah smoking have been regularly reviewed, updated and summarised over the last decade [3, 27]. CO levels are generally high, as early established by Sajid et al in 1993 [28]. However, the diverse types of charcoal, tobacco-based mixtures, the size of the device and, above all, the degree of ventilation (indoors, outdoors), play an important role in variations [3, 28]. Expired mean CO levels in smokers may thus reach values as low as 14.2 ppm in Jordan [29] and as high as 38.5 ppm at the end of a session (with quick-lighting charcoal) in a Lebanese café where patrons smoked cigarettes and hookah [30]. Levels are even more elevated in some ill-ventilated hookah lounges in France [3]. A study by Al-Kubati et al. on heart rate in hookah smokers, cited in a recent Cochrane review, did not assume that carbon monoxide levels might, perhaps, also account for observed differences with cigarettes, rather than the nicotine itself only [31, 32]. Besides, only 5 grams of the smoking product was used for a whole 45 minute session when the common value retained in most reliable studies of the Middle East are 20 grams [29, 30]. In these hard experimental conditions, the smoking mixture may char rapidly with a consequent overproduction of CO, carbonaceous material and tar [33, 34]. A recent study involving machine smoking shows that hookah mainstream smoke CO concentration is up to 13 times inferior than in cigarettes: 1.79 mg for a 1000 ml hookah (machine) puff and 1.06 mg for a 45 ml cigarette puff [35]. Another area of public health concern in the context of the global epidemic of hookah smoking is ETS (Environmental Tobacco Smoke). There has been a serious debate over statistics about cigarette ETS and their interpretation [36]. However, and in striking contrast with cigarettes, hookah does generate almost no side-stream smoke because of its peculiarities (charcoal topping the bowl and less elevated temperatures). So, the only smoke that should be taken into account is the one rejected by the smoker, i.e. the one filtered by the hookah at the level of the bowl, inside the water, along the hose and then by the smoker's lungs themselves. Consequently, the resulting smoke is expected to be less toxic for non-smokers than cigarette side-stream smoke. Notably, a great proportion of irritants, mainly aldehydes and phenols, are removed [19]. A team led by Guillerm in France early found that when passed through water (50 cm3), the combustion gases of cigarette smoke have no inhibitory effect on the respiratory epithelium cilia. The researchers concluded that narghile users can, "without apparent disorders, smoke dramatically greater quantities of tobacco than ours in our countries" [37]. Wynder et al have established that water filtered cigarette smoke is less toxic to clam gill tissue and that "a flask containing 200 ml of water dramatically can reduce the dose of ciliatoxic agents delivered to the ciliated epithelium" [38]. Weiss also reported that the effect of bubbling tobacco smoke through 15 ml of water was "equivalent to the effect of the better charcoal filters" [39]. Zaga and Gattavecchia have shown that the water in the vase of a hookah acts as an antioxydant against some short half-life free radicals [40]. Other substances are supposed to be affected by the water obstacle because of their solubility or the low temperatures: e.g. HCN, nitric oxides, etc. As for particles in the mainstream smoke, and particularly ultra fine ones (0.02 to 1 μm), a recent study shows that hookah smoke is up to 3 times less concentrated than cigarette smoke: 74.4 10⁹ for a 1000 ml hookah (machine) puff and 9.24 10⁹ for a 45 ml cigarette "puff" [35]. Similarly to ETS, there is a serious debate on the so-called "nicotine addiction" as this alkaloid is not seen as the central substance involved in the dependence phenomenon [41, 42]. Also, it is not the most dangerous one. Hookah dependence is very specific and research on it will help reconsider the "nicotine addiction" hypothesis. Recently, a Lebanese team has found that more than 90% of so-called "mild smokers" (3 pipes or less per week) and about 50% of the so-called "moderate" ones (3 to 6 pipes per week) are considered as non dependent [43]. A certain confusion is also a direct result of the misuse of smoking machines [33, 34]. A team in Kuwait has established with a rigorous methodology that the nicotine intake is not as high as in cigarettes [44]. Some of the smokers in our study, particularly the "heavy" ones, were obviously dependent. Further research is needed in this field and it is certainly too early to suggest the use of Nicotine "Replacement" Therapies and products to "hookah addicts", bupropion [32] or even Varenicline produced by Pfizer laboratories and marketed as Chantix and Champix.

Specific aspects of hookah smoking in Pakistan

The range of "normal" values provided by the manufacturer of our kits (DPC, USA) were: 2.1–6.2 ng/ml for male smokers (153 individuals study) and 1.3–4.9 ng/ml for female smokers (81 individuals study). The overall range of values (males+females) in these individuals was 1.3–6.2 ng/ml. The range was 1.1–3.2 ng/ml and 0.8–2.5 ng/ml for male (226 persons study) and female non-smokers respectively (262 persons study). The overall range (males+females) of values in non-smokers was 0.8 – 3.2.

In our last study we observed a mean CEA level of: 9.19 ng/ml (± 14.9) in 122 cigarette smokers; 7.16 ± 10.38 in mixed hookah smokers; and 2.35 ng/ml (± 0.71) in non-smokers [19]. Regarding levels in cigarette smokers, the chemiluminescent immunoassay, more sensitive and specific than the technique (Hansen-Z-gel) used by Alexander et al [11], may account for the differences. However, the origin and quality of tobacco and the environment should also be taken into consideration.

In our last study, CEA levels increased with the number of cigarettes smoked per day and the highest levels were reached by users who smoked more than 31 cigarettes per day [19]. Doll and Peto had established that the annual lung cancer incidence is: 0.273 × 10^-12. (cigarettes/day+6)².(age-22.5)^4.5 [45]. However, some decades ago, people used to smoke at a later age than today. The formula was revisited by Hill who insisted that the most relevant parameter for the assessment of tobacco consumption is duration and not dose (expressed in pack-years) [46]. If consumption is doubled, the risk is multiplied by 2. In striking contrast, if the duration is increased twofold, the risk is multiplied by 23 (2^4.5 exactly). Therefore, the exposure duration to tobacco smoke is much more important than the daily number of cigarettes. One conclusion is that quitting as early as possible remains the most powerful factor. Although critical, Lebeau also considers that there is no treshhold dose for risk [47].

In the case of the world fashionable tobacco-molasses smoking mixture called moassel (tobamel), users feel the smoke is very mild (because of the actual water trapping of notorious irritants such as aldehydes) and one direct consequence is that they often inhale considerable amounts of the smoke: randomly varying between 100 ml at least (but less sometimes) and up to 500 ml and sometimes more. These quantities of smoke go directly into their lungs with no previous dilution and stocking inside the mouth (as in the case of cigarettes)[3]. However, in the case of the Desi Punjab tobacco-molasses mixture smoked in Pakistan, and probably because of the different composition (no glycerol, more nicotine) and the higher temperatures, the smoke is not inhaled directly into the lungs and therefore it is diluted. Zahran stated that 15 minutes of its smoking "would provide approximately the same amount of tobacco smoke as one cigarette"[20].

A careful data analysis of our samples shows that CEA values are not normally distributed among the individuals of the various groups (Figures 6, 7, 8, 9, 10). We have therefore applied Wilconson's test to compare CEA values among the different groups. The p-values we obtained this way lead us to almost the same conclusions as those we get to with the t-test except for the comparison between overall CEA levels and non-smokers, where we get a non-significant result in Wilcoxon's test in contrast to a significant one in t-test. This may be due to the high sensitivity of Wilcoxon test over t-test. We are not getting a bimodal pattern in the distribution curve for our smokers, so the grouping according to such a pattern cannot be strictly followed. Therefore, the grouping on the basis of smoking level remains the good criterion for the grouping of our subjects.

The CEA test, although widely used for diagnosis and monitoring the therapy of cancer is not 100% specific for this disease [48]. So when someone's CEA level is high, it does not necessarily mean that the patient has cancer. Similarly, if someone's CEA level is low or close to lower normal limit, it does not mean that the individual is protected from cancer. In the light of these facts, we decided not to merge the light and medium groups of smokers.

Hookah smoking and Cancer

53 years ago, the British Medical Journal tried to answer the following question: "Does the custom of filtering tobacco smoke through water as in the Eastern hookah remove the noxious elements? Carcinoma of the lung is very rare, in my experience, in hookah-smoking Indians" [49]. A review of what research says about hookah smoking and cancer is therefore necessary to understand the possible influence of the type of exclusive hookah smoking ("light" and "medium" vs. "heavy") on CEA levels, supposed to reflect cancer risks. About half a century ago, Rakower and Fatal investigated this issue further to their analysis of rare available epidemiological statistics on narghile smoking [50]. However, the pioneer in this field is Angel Roffo in 1939 [51]. A recent review of his work was recently published although his very study on narghile was not included for some reason [52]. Roffo was the first to design a machine for the analysis of narghile smoke chemistry. He examined the fluorescence and spectrography of the tar filtered by both a water (narghile) and cotton filter. He was surprised by the filtration rate of the water itself: about 30%. However, he concluded that water and cotton filters could not be used as a means of "absolute prevention" for the prophylaxis of cancer. Lesions on animals appeared more lately than when the tar was directly applied on them. In the researcher's view, this represents a way to reduce the harm of tobacco use ("un medio de aminorar la accion del tabaquismo")[51]. Rakower and Fatal used a more sophisticated smoking machine and speculated on the reasons for which lung cancer would be less prevalent among narghile users: particularly the lower temperatures in relation to the formation of PAH (polynuclear aromatic hydrocarbons)[50]. This study has been sometimes cited in the available literature to support the statement that narghile smoking causes lung cancer [53], i.e. the opposite finding reached by its authors. This has resulted in a wide confusion [54]. More recently, a renowned Syrian lung specialist who has extensively studied narghile smokers, concluded that the low temperatures and filtration of part of the tar may account for the low rates of cancer observed in her country [55].

The hazards of tar, and particularly its carcinogenicity are directly related to the working temperatures whereby not only combustion and pyrolysis are involved, as a WHO report states, but also distillation, as emphasised by Baker et al [21, 56]. This is particularly true in the case of shisha smoking where the temperatures of the tobacco-molasses mixture in the bowl does not go in excess of 150°C, allowing a chemical reaction of the Maillard type [27]. Even when using a smoking machine and despite the bias these methods entail, the temperature hardly reaches 200°C [34, 35]. The more the temperature is elevated, the more carcinogenic the smoke is. In these conditions, hookah tar is qualitatively very different from that produced by cigarettes. Furthermore, in the case of the fashionable shisha (using flavoured molasses tobacco with glycerol), a great portion of the calculated "tar" is expected to be made up of glycerol which has proved not "adversely alter the smoke chemistry or biological effects normally associated with exposure to mainstream cigarette smoke" [57] as in the harm reduction Eclipse cigarette (about 40%). In any case, what might be really hazardous for tobacco smokers are, apart from PAH, nitrosamines the weight of which does not reach, in certain brands of cigarettes, the 10,000th part of the tar figures printed on the packets. Gray et al, in a study on Polish products, showed that cigarettes containing more nitrosamines were not those with a higher tar content [58]. Consequently, the rating of tar, particularly produced by smoking machines, makes no sense and may deceive tobacco users. Sound and deep research is needed in this field, all the more that a recent study by Sepetdjian et al, based on a smoking machine, has found great amounts of carcinogenic PAH [59]. The underlying methods, including the smoking topography, have been criticised for considerably reducing a highly complex human and social situation [34]. For example, the FTC (Federal Trade Commission) and ISO norms suggest the use of a 1 minute machine smoking interval between 2 puffs in the case of cigarettes for which the duration of a session barely exceeds 5 minutes. However, and by a striking contrast, the hookah smoking device used in the laboratory was based on steady puffs every 17 seconds. This implies that about 1 out of every 4 puffs is supposed to be a human breath for a whole one-hour session (171 regular puffs were drawn this way). Also, contrary to common practice in the real life, the charcoal was left in the same position over the bowl during all this period. In these conditions, the nature and yields of toxicants in the smoke are questionable. Furthermore, the low temperatures involved, as highlighted several times in our study [27, 33–35, 50, 55], do not theoretically allow for the abundant formation of hazardous PAHs. Sepetdjian et al suggest that one source for the PAHs might be the charcoal and that different types of the latter might induce different yields. As in the case of CO, Sajid et al had early established in 1993 that concentrations of this gas depend on the nature of charcoal (natural vs. commercial)[28]. In any case, assays on human subjects (urinary carcinogens, chemical or biological markers) would be more appropriate as in our previous study and the present one [19].

Diary of a Deputy Commissioner (DC)

In 1889 the British Deputy Commissioner (DC; Chief Controlling Officer of a district; mostly British national) in colonial India used to follow a small booklet written by an English DC of British times. The booklet contained instructions, which were compulsory for each DC to follow in order to control and frighten the public of their area. The instructions regarding number and kind of their personal staff were interesting. For each DC it was essential to keep following staff for his personal service at the expense of government.

Bera (table servant):1, Butler:1, Khansama (Cook):1–3, Servant:1, Sag Bardar (Dog carer):1, Saees (Horse carer):2, Massalchi (Massager), Hammal (porter; to carry the load):1, House maid:1, Hookah Bardar (hookah servant):1, Dhobi (Washerman):1, Darzi(tailor):1, Bahishti (Water carrier/sprinkler):1, Mali (gardener):1, Nai (Barber):1, Doodh wala (Milkman):1, Mehtar(sweeper):1, Pankha Quli (Fan operator);1, Patta Dar (Peons):3.

Queen used to rule India in those days. The empire was so big in those days that the sun did not set in the area occupied by the Great Britain. The D.Cs used to drink boiled water and eat the fruits dipped in antiseptic solutions. In order to avoid hot air they used to wear falalain (Flannel; special kind of cloth with soft and thin fiber) and used hand gloves and long shoes to avoid the mosquitoes.

In the brigade of their servants, hookah servant had a special rank. The cigarettes and cigars were not common in those days. However pipes were common but where East India Company left India for queen it also transferred hierarchy of hookah. Hookah smoking was very common in the beginning of 20^th century. English officers used to keep very decorated and attractive hookahs. The water used in these hookahs was mixed with flower extracts having very pleasant smell and the officers used to smoke in a retiring phase (laid on bed or seat of his vehicle). A servant used to handle the pipe of hookah during this period. He also used to blow the burning coal through a copper pipe (phoonkni). All officers used to keep these hookahs with them in lunch and dinner parties. The hookah carriers also accompanied him. After each meal of such parties a procession of hookah carriers used to come to the party venue-carrying hookahs to the party place. Every hookah holder used to take his hookah to his lord and used to stay respectfully for indefinite periods. The hookah sittings were full of manners and to pass over one man's pipe of hookah to the other was thought to be a serious disrespect. The maims (madams; respected ladies; wives of officers) had also the habit of smoking hookah. They used to wave/wrap the long pipes of hookah around their waists like snakes and thus enjoyed smoking. Ilaichi (cardamom; a special kind of seed in India with pleasant smell; used in herbal medicine), saffron and gold leaves (very thin sheets of gold) were mixed with tobacco. The mothers of these maims used to tell very proudly to their neighbors in Europe saying "Our daughters play with snakes and inhale the gold. When the wife of a deputy commissioner wanted to honor someone she allowed him/her to take few puffs of her hookah. On the other hand wife of a session judge wanted to compete her and exhibited her hookah frequently.

Other historical references, including translation from Urdu into English by KM Sajid

The Eighteenth century poetry reveals that most of the Indian high class liked hookah the most. The barons used to appoint poets to please them through their lyric creations. Nasikh was one of them. He says when admiring his lord:

Hookah jo hay Huzoor-e-Mualla kay hath mein

Goya keh kehkshan hay surrayya ke hath mein

Yeh sab baja hay laikin ae Nasikh tu arz kar

Bay jan bolta hay masiha kay hath mein.